Abstract: With characteristics of simple structure，high precision，and high operability，a gauss function model of predicting deformation data for dams is set up to overcome the shortcomings of the existing methods.The main function of ga（) in the genetic algorithm toolbox(MATLAB) is adopted to calculate the initial value of model coefficient，and then the fminsearch function in search toolbox is used to obtain the final value of model coefficient.16 sets of observation data are used to make modeling，and 4 groups of deformation data are used to test the model.The testing results show that the model has high accuracy，and its prediction is far better than other existing prediction model.This new model provides a new method and new idea for deformation prediction.

Key words: Deformation prediction, Gaussian function, Genetic algorithm, MATLAB